Cover Page
The handle http://hdl.handle.net/1887/45885 holds various files of this Leiden University dissertation.
Author: Kersten, K.
Title: Pulling the strings on anti-cancer immunity
Issue Date: 2017-02-07
Pulling the strings on anti-cancer immunity
Kelly Kersten
Based on the research described in this thesis and inspired by the song ‘Master of Puppets’ by Metallica, I envision cancer as a puppet-master restraining the protective function of the immune system. Part of my PhD work has focused on how tumor cells manipulate the function of immune cells to favor their spread throughout the body. In other words, cancer is pulling the strings on anti-cancer immunity to prevent destruction by the immune system. With the recent advances of combinatorial anti-cancer therapies (including immunomodulatory drugs) we can gain back control over the strings on anti-cancer immunity.
Cover design: Kelly Kersten & Tomasz Ahrends Artwork: Tomasz Ahrends
Lay-out: Jasper Koning (koningjj@gmail.com) Printing: Gildeprint, Enschede
ISBN: 978-94-6233-496-0
The printing of the thesis was financially supported by the NKI-AVL.
©2016 by Kelly Kersten. All rights reserved. No part of this thesis may be reproduced, stored in a retrieval system, or transmitted in any form or by any means without prior permission of the author and the publisher holding the copyright of the articles.
The research described in this thesis was performed at the Divisions of Molecular Biology and Immunology of the Netherlands Cancer Institute – Antoni van Leeuwenhoek Hospital (NKI-AVL), Amsterdam, The Netherlands, and was supported by the Dutch Cancer Society (KWF2011-5004) and the European Research Council (InflaMet 615300).
Pulling the strings on anti-cancer immunity
Proefschrift ter verkrijging van
de graad van Doctor aan de Universiteit Leiden op gezag van Rector Magnificus Prof.Mr. C.J.J.M. Stolker,
volgens besluit van het College voor Promoties te verdedigen op dinsdag 7 februari 2017
klokke 15.00 uur.
door Kelly Kersten geboren te Ede
in 1987
Promotores:
Prof.dr. Jos Jonkers, promotor
Dr. Karin E. de Visser, NKI-AVL, co-promotor Promotiecommissie:
Prof.dr. Joke Bouwstra, voorzitter Prof.dr. Meindert Danhof, secretaris Prof.dr. Ferry Ossendorp
Prof.dr. Bob van de Water
Prof.dr. Jacco van Rheenen (Utrecht University) Prof.dr. Jannie Borst (University of Amsterdam)
Chapter 1 Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7 Appendices
Scope of thesis
Genetically engineered mouse models in oncology and cancer medicine
EMBO Molecular Medicine, in revision.
IL-17-producing γδ T cells and neutrophils conspire to promote breast cancer metastasis
Nature 522:345-348 (2015)
Mammary tumor-derived CCL2 enhances pro-metastatic systemic inflammation through upregulation of macrophage-derived IL1β In revision.
Exploiting the immunomodulatory properties of chemotherapeutic drugs to improve the success of cancer immunotherapy
Frontiers in Immunology 6:516 (2015)
Dual immune checkpoint blockade synergizes with chemotherapy in a drug-dependent manner in a mouse model for de novo mammary tumorigenesis
In preparation General discussion English summary
Nederlandse samenvatting Curriculum Vitae
List of publications
9 19
43
71
97
125
151 171173 177179
Scope of thesis -1-
1
Breast cancer is the most common type of cancer in women; about 1 in 8 women will develop breast cancer during the course of her life 1. Moreover, breast cancer is the main cause of cancer-related mortality among women. The majority of these deaths is caused by metastatic disease which is still largely unexplored, poorly understood, and incurable 2. Most anti-cancer treatment strategies used to date are developed to target the primary tumor. However, we need to appreciate the fact that cancer is a systemic disease, and treatment of the primary tumor is often not sufficient to cure cancer patients.
Tumors do not merely consist of cancer cells, but together with a variety of stromal cell types like fibroblasts, vascular and lymphatic endothelial cells and infiltrating immune cells, form an entity collectively termed the tumor microenvironment (TME). In the past few decades it has become clear that the tumor microenvironment plays an important role in cancer development, progression and therapy responsiveness 3. Immune cells are of particular interest because of their paradoxical role in cancer progression and metastasis.
Cancer metastasis is a step-by-step process
The complexity of metastasis lies in its multistep nature. During primary tumor growth genetic alterations accumulate in cancer cells that allow their dissemination from the primary tumor. During dissemination, these cells have to cross multiple barriers like the basement membrane and extracellular matrix before invading surrounding tissues. Via a process called intravasation, disseminated cancer cells enter the blood stream and lymphatics. Once in the circulation, many cancer cells are cleared due to the high sheer stress and attack by the immune system. However, a minor fraction of surviving cancer cells can get trapped in the small capillary structures at distant sites and extravasate into the tissue. Here cancer cells can form micrometastatic lesions that sometimes remain dormant for long periods of time 4. However, when disseminated cancer cells reside in a permissive microenvironment, small lesions can progress to colonize distant organs forming macrometastatic disease.
During every step of the metastatic cascade there is a complex crosstalk between disseminated cancer cells and their surrounding microenvironment. As early as 1889 the English surgeon Stephen Paget proposed his ‘seed and soil’ hypothesis which states that metastasis depends on crosstalk between selected cancer cells (the ‘seeds’) and a specific organ microenvironment (the ‘soil’) 5. Only if these cancer cells end up in a supportive environment or niche they are able to survive and give rise to metastatic lesions. Emerging evidence indicates that the immune system plays an important role in priming the ‘soil’ for metastasis 6,7.
The paradoxical role of the immune system in cancer progression
The mammalian immune system consists of two arms that together help to protect the body from disease-causing infectious agents. The innate immune system — composed of monocytes, macrophages, neutrophils, dendritic cells, natural killer cells and mast cells — acts as a first line of defense and can rapidly eradicate invading pathogens. T and B cells compose the adaptive immune system and provide antigen-specific responses upon
encounter with a pathogen. Moreover, adaptive immune cells can provide immunological memory. Many cancers are characterized by the influx of large numbers of immune cells.
Recent studies show that the immune composition is predictive of prognosis in several cancer types 8,9. However, the functional role of the immune system in cancer progression is paradoxical; some immune cell populations harbor pro-tumorigenic properties, while other populations counteract tumorigenesis 3,10,11.
To mount effective anti-tumor immunity, tumor-associated antigens need to be taken up and processed by antigen-presenting cells, like dendritic cells. After receiving maturation signals, these cells migrate to the tumor-draining lymph nodes where the antigen is presented to naïve T cells. Upon activation, these tumor antigen-specific T cells migrate to the tumor bed to exert their cytotoxic function and eliminate cancer cells. Unfortunately, tumors elicit a variety of mechanisms to evade anti-tumor immunity and prevent destruction by the immune system, such as antigen-loss and dysfunctional T cell priming 12. In addition, many types of cancer are characterized by chronic inflammation which is one of the hallmarks of cancer 3,13. During chronic inflammation, tumor cells and inflammatory cells produce a variety of cytokines, chemokines and growth factors that favor the recruitment and polarization of immune cells, and induce angiogenesis and tissue remodeling. Moreover, inflammation often results in immunosuppression which is unfavorable for anti-tumor T cell responses. The complex reciprocal interactions between neoplastic cells and adaptive and innate immune cells create a delicate balance between pro- and anti-tumor immunity.
Immunotherapy as a therapeutic strategy to combat cancer
In the past years, cancer immunotherapy – harnessing the patient’s immune system to fight cancer – has proved to be a promising therapeutic strategy for several types of cancer
14. A growing body of data reports beneficial responses in predominantly immunogenic cancer types like advanced melanoma and lung cancer 15–18. However, a large proportion of patients does not show clinical benefit from cancer immunotherapy. Therefore the current focus in research is to better understand the underlying mechanisms of tumor- induced immune evasion to identify biomarkers that can predict whether a specific cancer patient will or will not respond to this type of therapy, and ultimately to develop strategies to overcome immune evasion.
It is now widely accepted that successful eradication of (metastasized) cancer requires a multi-disciplinary approach in which different anti-cancer treatment modalities are combined. While conventional therapies like chemotherapy, irradiation and targeted therapy usually show fast anti-tumor responses, the onset of acquired resistance often results in disease recurrence. In contrast, a proportion of patients treated with immunotherapy show slow but long-term durable anti-tumor immune responses, which makes immunotherapy an interesting modality to be combined with conventional anti- cancer therapies. To find the most optimal treatment combinations per cancer type, research in preclinical mouse cancer models is essential.
The research described in this thesis aims to gain a better understanding of the role of the immune system in cancer development and metastasis formation using preclinical
1
mouse models of metastatic breast cancer. With this knowledge, we aim to contribute to the development of immunomodulatory strategies to fight metastatic breast cancer and to increase the efficacy of conventional anti-cancer therapies.
Description of the chapters in this thesis
Despite the successful validation of novel anti-cancer drugs in preclinical models, the majority of phase III clinical trials fails to meet their primary endpoint 19. The poor translation from preclinical mouse models to clinical practice illustrates the insufficient predictive power of the preclinical models that are currently used. To improve these disappointing statistics, it is desirable that preclinical models faithfully recapitulate human cancer. Genetically engineered mouse models (GEMMs) have proved indispensable for gaining biological insight into the many different aspects of human cancer, including genetic driver mutations, onset of metastasis, interaction with the surrounding microenvironment and responsiveness to anti-cancer therapies. Moreover, the presence of an intact immune system in these mice that co-evolves with de novo tumor development is very important in the context of studying the anti-cancer efficacy of immunomodulatory drugs. In Chapter 2 we propose how the current technological advances in mouse cancer model engineering can contribute to improve the predictive power of preclinical studies. Ultimately this will result in more effective anti-cancer treatment strategies.
Our research described in the first part of this thesis is aimed at gaining a better understanding of the role of the immune system in breast cancer progression and metastasis formation. In this work we made use of a GEMM for de novo mammary tumorigenesis; i.e. K14cre; Cdh1F/F;Trp53F/F mice 20. The mammary tumors that spontaneously develop in these animals closely resemble a subtype of human breast cancer known as invasive lobular carcinoma (ILC), which accounts for approximately 10%
of all breast cancer cases 21. We used this mouse model to study the tumor-induced mechanisms of immune evasion during breast cancer progression. In Chapter 3, we demonstrate that de novo mammary tumors that arise in the conditional K14cre;Cdh1F/
F;Trp53F/F mouse model, induce a systemic pro-inflammatory cascade to facilitate breast cancer metastasis to distant organs. This pro-metastatic inflammation is characterized by interleukin (IL)-17 expressing γδ T cells and the subsequent expansion and polarization of immunosuppressive neutrophils. These neutrophils actively suppress the activity of CD8+ T cells via iNOS. We found that IL-17 expression by γδ T cells is induced by mammary tumor-derived IL-1β. In Chapter 4 we report an additional regulator of this tumor- induced systemic inflammatory cascade. We identified CCL2, a chemokine that is highly expressed in K14cre; Cdh1F/F;Trp53F/F mammary tumors, as a key driver of the γδ T cell – IL17 – neutrophil axis by inducing IL-1β expression in tumor-associated macrophages.
In line with these findings, we show that expression of CCL2 positively correlates with IL1Β and macrophage markers in human breast tumors. Together our findings suggest that interfering with this pro-metastatic inflammatory cascade may provide therapeutic options for patients with metastasized breast cancer.
In the second part of this thesis I focused on the role of the immune system in therapy responsiveness. Studies suggest that the efficacy of anti-cancer therapy is (in part) dependent on immune-mediated mechanisms 22,23. The success of cancer immunotherapy has reinvigorated the search for combinatorial treatment strategies that induce cancer cell death and boost anti-tumor immunity to optimize therapeutic response rates. One of the strategies proposed in this thesis is to combine the treatment of immunotherapy with conventional chemotherapy. Several chemotherapeutics have immunomodulatory properties and affect different populations of immunosuppressive immune cells. For example, cyclophosphamide (when administered in low doses) targets regulatory T cells 24, and gemcitabine specifically targets myeloid-derived suppressor cells (MDSC) that counteract T cell activity 25. In Chapter 5, we summarize preclinical and clinical data that support the notion that combining T cell boosting immune checkpoint inhibitors with conventional chemotherapeutics that alleviate immunosuppression will enhance the efficacy of anti-cancer treatment strategies for patients.
Although the presence of tumor-infiltrating lymphocytes correlates with a good prognosis 26,27, breast cancer is not considered a highly immunogenic type of cancer.
Clinical trials are currently ongoing to explore the efficacy of immunotherapy to enhance tumor-reactive T cells in breast cancer patients. Since objective response rates presented so far range from 5–20% 28–31, a substantial fraction of breast cancer patients requires optimized combinatorial treatment approaches. Chapter 6 describes our research in which we utilized the K14cre;Cdh1F/F;Trp53F/F mouse model to explore the applicability of immunotherapy by immune checkpoint blockade in spontaneous breast cancer. We found that dual immune checkpoint blockade with anti-PD-1 and anti-CTLA-4 does not improve tumor-specific survival of mice. However, when combined with conventional chemotherapy we find synergistic responses in a drug-dependent manner. Improved anti-tumor responses were dependent on CD8+ T cells. These results have important implications for treatment strategies in the clinic, because it shows the importance of the chemotherapy of choice. More importantly, our results demonstrate that – even in relatively poorly immunogenic cancer types – a combination of chemo- and immunotherapy is able to unleash anti-tumor immunity to combat cancer.
Chapter 7 summarizes the main results described in this thesis and puts these findings in context of the current literature. It also provides suggestions for clinical implications and future directions.
Taken together, the preclinical research described in this thesis demonstrates that anti- tumor immune responses occur, but are overruled by tumor-induced immune-evading mechanisms in a mouse model of breast cancer. By inducing a systemic inflammatory and immunosuppressive state the tumor manipulates the function of immune cells favoring its dissemination. In other words, the tumor is ‘pulling the strings on anti-cancer immunity’ to prevent destruction by the immune system. With the recent advances in the field of immunomodulatory drugs we now have the proper tools to overrule this systemic immune evasive state and gain back control over the strings on anti-cancer immunity.
Ultimately, this will improve therapeutic strategies and improve cancer patient care.
1
References
1. www.cijfersoverkanker.nl.
2. Weigelt, B., Peterse, J. L. & van’t Veer, L. J. Breast cancer metastasis: markers and models. Nat. Rev.
Cancer 5, 591–602 (2005).
3. Hanahan, D. & Coussens, L. M. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell 21, 309–322 (2012).
4. Ghajar, C. M. Metastasis prevention by targeting the dormant niche. Nat. Rev. Cancer 15, 238–247 (2015).
5. Paget, S. The distribution of secondary growths in cancer of the breast. Lancet 1, 571–573 (1889).
6. McAllister, S. S. & Weinberg, R. A. The tumour-induced systemic environment as a critical regulator of cancer progression and metastasis. Nat. Cell Biol. 16, 717–727 (2014).
7. Kitamura, T., Qian, B.-Z. & Pollard, J. W. Immune cell promotion of metastasis. Nat. Rev. Immunol. 15, 73–86 (2015).
8. Gentles, A. J. et al. The prognostic landscape of genes and infiltrating immune cells across human cancers. Nat. Med. 21, 938–945 (2015).
9. Mlecnik, B. et al. Integrative Analyses of Colorectal Cancer Show Immunoscore Is a Stronger Predictor of Patient Survival Than Microsatellite Instability. Immunity 44, 698–711 (2016).
10. de Visser, K. E., Eichten, A. & Coussens, L. M. Paradoxical roles of the immune system during cancer development. Nat. Rev. Cancer 6, 24–37 (2006).
11. DeNardo, D. G. & Coussens, L. M. Inflammation and breast cancer. Balancing immune response: crosstalk between adaptive and innate immune cells during breast cancer progression. Breast Cancer Res. 9, 212 (2007).
12. Chen, D. S. & Mellman, I. Oncology meets immunology: the cancer-immunity cycle. Immunity 39, 1–10 (2013).
13. Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 144, 646–674 (2011).
14. Couzin-Frankel, J. Cancer Immunotherapy. Science 342, 1432–1433 (2013).
15. Hodi, F. S. et al. Improved survival with ipilimumab in patients with metastatic melanoma. N. Engl. J.
Med. 363, 711–723 (2010).
16. Brahmer, J. R. et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N. Engl.
J. Med. 366, 2455–2465 (2012).
17. Topalian, S. L. et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N. Engl. J.
Med. 366, 2443–2454 (2012).
18. Wolchok, J. D. et al. Nivolumab plus ipilimumab in advanced melanoma. N. Engl. J. Med. 369, 122–133 (2013).
19. Reichert, J. M. & Wenger, J. B. Development trends for new cancer therapeutics and vaccines. Drug Discov. Today 13, 30–37 (2008).
20. Derksen, P. W. B. et al. Somatic inactivation of E-cadherin and p53 in mice leads to metastatic lobular mammary carcinoma through induction of anoikis resistance and angiogenesis. Cancer Cell 10, 437–449 (2006).
21. Arpino, G., Bardou, V. J., Clark, G. M. & Elledge, R. M. Infiltrating lobular carcinoma of the breast: tumor characteristics and clinical outcome. Breast Cancer Res. 6, R149-156 (2004).
22. Zitvogel, L., Galluzzi, L., Smyth, M. J. & Kroemer, G. Mechanism of action of conventional and targeted anticancer therapies: reinstating immunosurveillance. Immunity 39, 74–88 (2013).
23. Coffelt, S. B. & de Visser, K. E. Immune-mediated mechanisms influencing the efficacy of anticancer therapies. Trends Immunol. 36, 198–216 (2015).
24. Ghiringhelli, F. et al. Metronomic cyclophosphamide regimen selectively depletes CD4+CD25+
regulatory T cells and restores T and NK effector functions in end stage cancer patients. Cancer Immunol.
Immunother. 56, 641–648 (2007).
25. Suzuki, E., Kapoor, V., Jassar, A. S., Kaiser, L. R. & Albelda, S. M. Gemcitabine selectively eliminates splenic Gr-1+/CD11b+ myeloid suppressor cells in tumor-bearing animals and enhances antitumor immune activity. Clin. Cancer Res. 11, 6713–6721 (2005).
26. Mahmoud, S. M. A. et al. Tumor-infiltrating CD8+ lymphocytes predict clinical outcome in breast cancer.
J. Clin. Oncol. 29, 1949–1955 (2011).
27. Ali, H. R. et al. Association between CD8+ T-cell infiltration and breast cancer survival in 12,439 patients.
Ann. Oncol. 25, 1536–1543 (2014).
28. Emens, L. A. et al. Abstract PD1-6: Inhibition of PD-L1 by MPDL3280A leads to clinical activity in patients
with metastatic triple-negative breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA):
AACR; Cancer Res. 75, Abstract nr PD1-6 (2015).
29. Dirix, L. Y. et al. Abstract S1-04: Avelumab (MSB0010718C), an anti-PD-L1 antibody, in patients with locally advanced or metastatic breast cancer: A phase Ib JAVELIN solid tumor trial [abstract]. In:
Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res. 76, Abstract nr S1-04 (2016).
30. Rugo, H. S. et al. Preliminary efficacy and safety of pembrolizumab (MK-3475) in patients with PD- L1–positive, estrogen receptor-positive (ER+)/HER2-negative advanced breast cancer enrolled in KEYNOTE-028. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res. 76, Abstract nr S5-07 (2016).
31. Nanda, R. et al. Pembrolizumab in Patients With Advanced Triple-Negative Breast Cancer: Phase Ib KEYNOTE-012 Study. J. Clin. Oncol. (2016). doi:10.1200/JCO.2015.64.8931
-2-
Genetically engineered mouse models in oncology and cancer medicine
Kelly Kersten1, Karin E. de Visser1, Martine H. van Miltenburg2, Jos Jonkers2
1Division of Immunology and 2Division of Molecular Pathology, Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands
EMBO Molecular Medicine, in revision.
Abstract
Genetically engineered mouse models (GEMMs) have made significant contributions to the field of cancer research. Tissue-specific induction of defined driver mutations in GEMMs triggers development of tumors in a natural immune-proficient microenvironment.
These tumors closely mimic histopathological and molecular features of their human counterparts, and display genetic heterogeneity, thus faithfully recapitulating the natural course of human cancer. GEMMs capture both tumor cell-intrinsic and -extrinsic factors that drive de novo formation of tumors and progression toward metastatic disease, and are therefore indispensable for preclinical research. GEMMs have successfully been used to validate candidate cancer genes and drug targets, assess therapy efficacy, and evaluate mechanisms of drug resistance. Great efforts are made to further fine-tune engineering of GEMMs and to align in vivo preclinical testing in advanced mouse models with clinical studies in patients, which is anticipated to speed up the development of novel therapeutic strategies and their translation into the clinic.
Pending issues
• Understanding of tumor cell-intrinsic and -extrinsic mechanisms underlying cancer and metastasis development, and therapy resistance.
• Development of multidisciplinary therapeutic strategies including conventional anti-cancer drugs and immunotherapy to successfully fight disseminated cancer.
• Reduction of time and costs to generate next-generation genetically engineered mouse models that closely recapitulate human cancer.
• Close alignment of preclinical mouse studies and human clinical trials to improve cancer patient care.
2
Introduction
Despite the fact that survival rates of cancer patients have improved over the last decades, we are still facing numerous challenges in the clinic. One of the major problems is the development of drug resistance. Monotherapy with targeted anti- cancer agents or chemotherapeutics can result in drug resistance caused by de novo mutations or outgrowth of pre-existing therapy-resistant clones within heterogeneous tumors. Moreover, after seemingly successful treatment, small numbers of drug-tolerant tumor cells can survive treatment and remain dormant for extended periods of time and eventually relapse to form recurrent disease that can be phenotypically different from the original tumor 1,2. Another major challenge is metastatic disease, which accounts for over ninety percent of cancer-related deaths 3. These secondary tumors are often unresponsive to therapy and are at present mostly incurable. Encouraging advancements have been made with cancer immunotherapy, aimed at harnessing the patient’s immune system to attack cancer. However, even though long term durable responses are observed in some cases, a large proportion of cancer patients does not show clinical benefit 4.
Successful treatment of cancer requires a multidisciplinary approach in which different strategies, such as surgery, irradiation, cytotoxic therapy and immunotherapy, are combined. In order to design such combinations, it is critical to improve our insights into the cancer cell-intrinsic and -extrinsic mechanisms underlying tumor development and metastasis, and therapy responsiveness. To find the most efficacious treatment for different cancer types, we heavily rely on preclinical research in animal models.
Despite successful validation of novel anti-cancer therapies in conventional preclinical mouse models based on xenotransplantation of established human cancer cell lines or allotransplantation of mouse tumor cell lines, the majority of the phase 3 clinical trials fail 5. The overall poor clinical predictability of these conventional in vivo tumor models emphasizes the need for more advanced preclinical in vivo models with a better predictive power. Until fairly recently, progress in the field was hampered by the poor availability of preclinical models that closely recapitulate the natural course of human cancer. However, recent technological developments have led to fast track generation of sophisticated mouse models that more closely mimic human cancer in terms of genetic composition, interactions of cancer cells with their tumor microenvironment, drug response and resistance. These next generation genetically engineered mouse models (GEMMs) are of great importance to improve our understanding of the complex mechanisms underlying cancer biology, and are anticipated to improve translation of new therapeutic strategies into the clinic — ultimately leading to increased survival of cancer patients. This review describes the evolution and recent technological advances of mouse model engineering, and the applications of the resulting models in basic and translational oncology research.
Evolution of mouse cancer modeling
Over the years, novel advances in the field of genome editing have led to the generation of various mouse models to study cancer biology. Here, we give a historic overview of the development of mouse models that are mostly used in cancer research.
Cancer cell line transplantation models
Allograft and xenograft cell line transplantation mouse models are the most commonly used mouse models, as they allow for rapid testing of potential cancer and metastasis-related genes and are often used for preclinical drug testing. Moreover, these cells – when tagged with biomarkers such as luciferase or fluorescent proteins to allow non-invasive imaging – have proven informative to identify metastasis-related genes. For example, orthotopic and intravenous administration of breast cancer cells has shed light on the mechanisms underlying organ tropism and metastatic dormancy 6–9. Nevertheless, as cancer cell lines contain multiple mutations from the start and acquire additional aberrations when cultured in 2D for extended periods of time, these inoculation models do not reflect the morphology and genetic heterogeneity of human cancers, and are therefore commonly poor predictors of clinical response. While allografting of mouse cancer cell lines can be performed in immune proficient hosts, xenotransplantation of cell lines must be performed in immunocompromised mice to prevent rejection, which makes them less suitable to study the roles of the immune system in tumor development and therapy response.
Patient-derived tumor xenografts
Patient-derived tumor xenograft (PDTX) models are derived from fresh human tumor biopsies that are transplanted in immunodeficient mice. Unlike cell line transplantation models, PDTX tumors maintain the molecular, genetic and histological heterogeneity as observed in cancer patients, even after serial passaging in mice 10. Therefore, PDTX models can be valuable tools to define personalized medicine as was demonstrated by preclinical drug screening in PDTX models of non-small cell lung cancer (NSCLC) 11–13, breast cancer 14, melanoma 15,16, prostate cancer 17,18 and colorectal cancer 19–24. High- throughput efforts are now undertaken using PDTX models to predict responses of clinical drug candidates. Approximately 1000 PDTX models were established with a diverse panel of mutations, and subsequently used for in vivo compound screens, yielding correlations between drug response and tumor genotype that were both reproducible and clinically translatable 25. In a recent study using PDTX models of triple-negative breast cancer, single-cell gene expression analysis revealed that early stage metastatic cells express distinct signatures enriched in stem-like genes, identifying novel potential drug targets to tackle metastatic breast cancer 26.
Unfortunately, a major obstacle of PDTX modeling is the disappointing take rate of various tumor types, such as estrogen receptor-positive breast cancer and prostate cancer 27,28. In addition, PDTX modeling must be performed in immunocompromised mice, thereby circumventing the natural anti- and pro-tumor activity provided by the
2
adaptive immune system. Given the complex crosstalk between adaptive immune components, the innate immune system and cancer cells, it is important to realize that PDTX models can provide clinically valuable data, albeit in the absence of the influential adaptive immune system. Current efforts to generate humanized mice by engrafting immunodeficient mice with human CD34+ hematopoietic stem cells or precursor cells have shown remarkable progress 29,30. Although reconstitution of immune cells from specific lineages remains challenging, the introduction of transgenes encoding human cytokines, chemokines and growth factors can support the development of human myeloid cells in mice. To support development of HLA-restricted T cells, recipient immunodeficient mice can be further optimized by transgenic expression of human HLA molecules and deficiency of mouse MHC class I and II molecules. While the limited availability of hematopoietic donor stem cells (obtained from umbilical cord blood or fetal liver) and the relatively high costs of these models are potential disadvantages, humanized mouse models could provide a useful platform for preclinical evaluation of immunotherapeutics.
Modeling de novo cancer in genetically engineered mice
In the early 1980s, the first cloned cancer genes were introduced into the genome of transgenic mice, which were termed oncomice 31. The first oncomouse was a GEMM with transgenic expression of a specific activated oncogene (v-HRas) under control of a mammary-specific promoter (MMTV), making the mouse prone to developing mammary tumors 32. The first oncomice led to great excitement in the cancer research community as they provided unambiguous proof for the hypothesis that oncogene expression in normal cells could lead to tumor formation 32–36. With the development of gene knockout technology in 1992, also cancer predisposition in tumor suppressor gene (TSG) knockout mice could be studied 37.
Though oncomice and TSG knockout mice have provided a wealth of knowledge, they also have their limitations. Given that transgenes are expressed in all cells of a particular tissue and TSGs in knockout mice are inactivated in all cells of the animal, these models fail to mimic sporadic cancers in which accumulation of genetic events in a single cell results in tumorigenesis in an otherwise healthy organ. To circumvent this, more sophisticated mouse models are currently available that allow somatic inactivation of tumor suppressors or activation of (mutant) oncogenes in conditional GEMMs 38. One of the first examples is the generation of a mouse colorectal cancer model using Cre- loxP mediated somatic inactivation of Apc. With this technique any gene flanked by loxP recombination sites will be deleted after activation of the Cre-recombinase. APC loss in intestinal epithelial cells was sporadically induced through adenovirus-mediated delivery of Cre-recombinase, resulting in the rapid onset of colorectal adenomas that shared many features with adenomas in familial adenomatous polyposis coli (FAP) patients 39. By introducing mutations associated with a specific type of cancer one can generate mouse models that closely mimic the histopathological, molecular and clinical features of tumors in patients 40,41.
Induction of somatic mutations at a chosen time and in a specific tissue can be achieved by using Cre-ERT fusion proteins, in which a mutated hormone-binding domain of the estrogen receptor (ERT) is fused to the Cre-recombinase. Cre-ERT is an inducible Cre-recombinase: administration of the estrogen analog tamoxifen leads to post-translational activation of Cre-recombinase activity and excision of the target gene flanked by loxP sites. Hence, mice with (tissue-specific) expression of Cre-ERT allow for spatiotemporally controlled Cre-mediated genomic recombination upon administration of tamoxifen 42.
Although the Cre-loxP system can be applied to alter the expression of more than one gene, it does so simultaneously, and therefore does not fully mimic the sequential accumulation of mutations during multistep carcinogenesis. Recently, an inducible dual-recombinase system was developed which combines Flp-FRT and Cre-loxP recombination systems, allowing sequential genetic manipulation of gene expression by two independent recombination systems 43. This approach allows for (i) independent targeting of tumor cell autonomous and non-autonomous pathways/processes, (ii) sequential induction of mutations to faithfully model human multistep carcinogenesis, and (iii) genetic validation of therapeutic targets in autochthonous tumors.
Mouse models to study oncogene addiction
Some tumors are highly dependent on a single oncogene for their growth, a phenomenon called ‘oncogene addiction’. Conditional GEMMs are unsuitable models to determine oncogene addiction, as the genetic lesion is irreversible, and thus requires another layer of regulation. Oncogene-ERT fusions can be employed to control oncogene expression;
for example, Trp53KI/KI mice in which both Trp53 alleles are replaced by the tamoxifen- inducible Trp53-ERT variant, have been used to determine the therapeutic efficacy of p53 restoration in established tumors 44.
Also systems for doxycycline-regulatable gene expression have been successfully used in GEMMs to turn oncogenes on, thereby allowing tumorigenesis; and off to investigate how established tumors respond to oncogene inactivation (43,44,45). To give an example, continuous expression of a doxycycline-inducible Myc transgene in hematopoietic cells resulted in the formation of malignant T cell lymphomas and acute myeloid leukemias that regressed upon de-induction of Myc expression 47. The long-term effects of temporal MYC de-induction seem to differ between cancer types. For example, brief inactivation of MYC in osteogenic sarcomas resulted in sustained regression due to differentiation of sarcoma cells into mature osteocytes 48. In contrast, invasive liver cancers regressed after MYC inactivation, but residual tumor cells remained dormant and immediately restored their neoplastic features upon MYC reactivation 49.
2
Speeding up and fine-tuning mouse cancer modeling
Although GEMMs have proven to be valuable tools for cancer research, there are still aspects that can be improved. A major limitation of germline GEMMs is that development and validation of these models is time-consuming, laborious and expensive. This is exemplified when a novel germline mutation has to be introduced in an existing multi- allelic mouse model, as this requires extensive breeding. The rapidly increasing number of mutations identified in cancer sequencing studies calls for novel mouse modeling strategies that enable accelerated in vivo evaluation of candidate cancer genes and patient-relevant allelic variants of known cancer genes.
Embryonic stem cell-based mouse cancer models
To speed up the generation of novel GEMMs of human cancer, embryonic stem cells (ESCs) can be genetically altered and used to produce cohorts of non-germline GEMMs 50. An alternative approach is the recently developed GEMM-ESC strategy, which employs ESCs that are derived from existing (multi-allelic) GEMMs. These GEMM-derived ESCs can be used for rapid introduction of additional genetic modifications and subsequent production of chimeric mice that show the same characteristics as the established GEMM but now contain the additional genetic modification 51,52.
In vivo RNA interference
RNA interference (RNAi) by short hairpin RNAs (shRNAs) allows reversible silencing of gene expression without modifying the genome, and therefore it can be used as an alternative to homologous recombination-based gene inactivation approaches. RNAi- based genetic screens have proven powerful tools to rapidly identify and validate cancer genes. In vivo RNAi screens have been successfully used to identify novel TSGs in mouse models of hepatocellular carcinoma and lymphoma 53–55, and to identify genes involved in resistance to the tyrosine kinase inhibitor sorafenib in liver cancer 56. Moreover, the development of systems for doxycycline-inducible shRNA expression in transgenic mice allows reversible expression of shRNAs in a time- and tissue-specific manner 57,58. Using the latter approach, Dow et al. have shown that shRNA-mediated APC suppression in the presence of Kras and Trp53 mutations induces intestinal carcinomas, which undergo sustained regression upon restoration of APC expression by turning off shRNA expression, highlighting the WNT pathway as a therapeutic target for treatment of colorectal cancer 59.
Genome editing using CRISPR/Cas9 technology
In the past decades, additional approaches for genome editing have been developed such as Zinc-finger nucleases (ZFNs) and transcription-activator-like effector nucleases (TALENs) 60,61. These approaches have now been outperformed by the development of CRISPR/Cas9 systems for genome editing 62, which have revolutionized biological research over the past three years and are considered the biggest game changer since
PCR. The CRISPR (clustered regularly-interspaced short palindromic repeats) – Cas9 system was first discovered as a prokaryotic immune system that confers resistance to foreign genetic elements, but soon thereafter has been exploited to achieve gene editing 63–65. By using appropriate single-guide RNAs (sgRNAs), the Cas9 nuclease can be directed to any genomic locus, where it induces double-stranded cleavage of matching target DNA sequences, leading to gene knockout 62. The CRISPR/Cas9 system can also be used to introduce defined mutations or loxP/FRT recombination sites, by simply co- introducing oligonucleotides that can serve as a template for repair of the Cas9-induced break 66.
CRISPR/Cas technology seems the system of choice for rapid cancer modeling in mice, as it has proven to be an efficient gene-targeting strategy with the potential for multiplexed genome editing 67. Virtually all (combinations of) genetic alterations found in human tumors can now be rapidly introduced in the mouse germline, including (conditional) gene deletions 68,69, point mutations 68 and translocations 70–72. Other groups have successfully used CRISPR/Cas9 technology for somatic editing of oncogenes and TSGs in mice. These efforts have led to a new generation of non-germline models of hepatocellular carcinoma 73,74, lung cancer 75,76, brain cancer 77, pancreatic cancer 78,79 and breast cancer 80.
The CRISPR/Cas9 system has recently been modified to induce target gene repression (CRISPRi) or activation (CRISPRa) 81. These modified systems may be used to generate mice with inducible and reversible activation of oncogenes and/or inactivation of TSGs.
Though extremely powerful, CRISPR/Cas9 based systems for in vivo gene editing may also have certain drawbacks. For example, current CRISPR/Cas9 strategies are not suited to validate the oncogenic potential of putative oncogenes. To this end, CRISPRa-based systems may be used to activate transcription of target genes 82. Moreover, somatic delivery of Cas9 may trigger Cas9-specific immune responses resulting in clearance of Cas9 expressing cells 80,83. To circumvent this issue, experiments should be performed in immunodeficient animals or mice that are engineered to develop tolerance to Cas9. Finally, CRISPR/Cas9 mediated genome editing may create unwanted off-target mutations that may be circumvented by employing mice with inducible expression of a Cas9n ‘nickase’ variant 84.
Fine-tuning mouse cancer modeling with patient-relevant alleles
Many cancer-predisposing germline mutations and somatic mutations in human TSGs are missense or nonsense mutations that may result in the production of a mutant or truncated protein with residual activity. Such mutations are not adequately modeled in (conditional) knockout mice, in which deletion of one or more exons leads to complete loss of the protein. It is therefore essential to generate mouse models carrying patient- relevant mutations to study their contribution to tumorigenesis and therapy response.
Several studies have shown that patient-relevant TSG mutations in mice induce different phenotypes compared to the null-alleles. Compared to Trp53 knockouts, patient- relevant Trp53 hotspot mutations in mice were shown to have enhanced oncogenic
2
activity 85,86. Similarly, introduction of patient-relevant Brca1 mutations in a conditional mouse model of BRCA1-associated breast cancer showed that, in contrast to Brca1- null tumors, mammary tumors with expression of Brca1 alleles harboring mutations in the RING domain readily acquired resistance to DNA-damaging drugs due to residual activity of the RING-less BRCA1 protein 87,88. Thus, by introducing specific somatic or germline mutations into GEMMs, the causal link between these mutations and therapy responsiveness can be determined.
Applications of GEMMs in basic cancer research
The generation of GEMMs has been detrimental for basic cancer research. Here, we discuss how GEMMs have contributed to understanding the basic intrinsic and extrinsic aspects of cancer biology.
Validation of candidate cancer genes
Given the growing number of candidate cancer genes that are identified in large-scale tumor sequencing studies, there is a clear need for rapid in vivo strategies to validate these genes. Considering their speed and relative simplicity, GEMM-ESC and CRISPR/
Cas technologies are the methods of choice for fast-track validation of candidate cancer genes. Especially non-germline models based on somatic CRISPR/Cas9-mediated gene editing enable in vivo validation of (combinations of) candidate cancer genes in a truly high-throughput manner, as was demonstrated in a mouse model for pancreatic cancer 79. Here, transfection-based multiplexed delivery of Cas9 and sgRNAs targeting 13 different cancer genes induced pancreatic cancer (PDAC) in the majority of mice.
The PDACs displayed genome editing of over 60% of the target genes, indicating clonal expansion of CRISPR/Cas9-induced driver mutations that induce cancer 79. Likewise, GEMMs with doxycycline-inducible Cas9 expression were employed to validate defined combinations of intestinal cancer genes, e.g. Apc and Trp53 84. Besides modifying TSGs, CRISPR/Cas9 technology can be applied to validate the oncogenicity of chromosomal rearrangements, such as the Eml4-Alk gene fusion observed in lung cancer 89.
Determining cells-of-origin of cancers
Identifying the cancer cell-of-origin may provide important information for the development of improved therapeutic strategies. Studies in GEMMs have successfully identified the cell-of-origin for several different cancer types. For example, the cell- of-origin of small cell lung cancer (SCLC) was determined by intra-tracheal injection of cell-type-restricted Adeno-Cre viruses, to inactivate Trp53 and Rb1 in Clara, neuro- endocrine (NE) and alveolar type 2 (SPC) cells, respectively. Trp53 and Rb1 inactivation in these specific cell types of the lung resulted in differences in tumor onset and tumor phenotype, and identified NE cells (and to a lesser extent SPC cells) as the cell-of-origin in SCLC 90. Cell-of-origin studies can also deliver surprising results, as was the case for BRCA1-related basal-like breast cancer. While BRCA1-related basal-like breast cancer
was previously postulated to originate from basal epithelial stem cells, cell-of-origin studies in GEMMs revealed that in fact luminal progenitors are the source of basal-like tumors 91. Genetic aberrations, such as Pik3ca mutations, can have a profound effect on the stem cell pool, as was demonstrated recently by two independent laboratories.
Expression of Pik3caH1047R was shown to evoke dedifferentiation of lineage-committed mammary epithelial cells into a multipotent stem-like state 92,93. Interestingly, the cell- of-origin of Pik3caH1047R mammary tumors dictates their malignancy, highlighting the importance of pinpointing the cancer cell-of-origin to improve specificity of anti-cancer drugs and therapeutic outcome.
Studying the contribution of the tumor microenvironment
GEMM models have been fundamental in deciphering the contribution of tumor cell- extrinsic factors such as cancer-associated fibroblasts (CAFs) and immune cells to tumorigenesis. CAFs are important cellular components of the tumor microenvironment as they regulate deposition of extracellular matrix (ECM) and formation of basement membrane by synthesizing ECM components such as collagen, fibronectin and laminin.
Moreover, fibroblasts are a source of various soluble mediators including matrix metalloproteases (MMPs), which enable ECM turnover, reinforcing their crucial role in maintaining ECM homeostasis 94. Studies in GEMMS have demonstrated dual roles of fibroblasts in cancer. During malignant transformation of epithelial cells, CAFs can stimulate tumor progression by enhancing inflammation, angiogenesis and ECM remodeling, as was demonstrated in the K14-HPV16 squamous skin cancer model 95. In contrast, a recent study demonstrated that genetic in vivo depletion of CAFs accelerates progression of pancreatic cancer 96, suggesting a tumor-restraining role for CAFs. The same controversy holds true for immune cells: originally it was hypothesized that immune cells suppress tumorigenesis by attacking transformed cells; however, work of recent years has revealed that these cells can also act as tumor-promoting entities. Early studies in the K14-HPV16 model have shown that mast cells and bone marrow-derived cells promote squamous skin cancer by activating angiogenesis and by reorganizing stromal architecture via MMP9 97,98. Using the same skin cancer model, chronic inflammation was found to promote de novo carcinogenesis in a B lymphocyte-dependent manner 99. Likewise, the tumor-promoting roles of tumor-associated macrophages (TAMs) 100,101 and neutrophils 102 have been described in several studies, emphasizing that immune cells can act as coconspirators in tumor development and progression.
Deciphering spontaneous metastasis formation
Despite the advancement of therapeutic options in the clinic, metastatic disease remains the primary cause of cancer-related death. The metastatic cascade is a complex multi-step process dictated by a constant crosstalk between cancer cells and their microenvironment 103,104. Most preclinical metastasis research has been performed in cell line inoculation models, which do not recapitulate the subsequent steps of the metastatic process as it occurs in patients. Spontaneously
2
metastasizing GEMMs provide unique opportunities to study metastasis because the entire cascade occurs de novo in a natural setting (Figure 1). A complication of the use of GEMMs for metastasis research is that these mice generally need to be sacrificed due to their primary tumor burden, before macroscopic metastases have developed. This problem can be overcome by orthotopic transplantation of GEMM- derived tumor fragments – which maintain the intratumoral heterogeneity of donor tumors – followed by surgical resection, allowing the development of clinically overt metastatic disease 105. GEMMs that closely recapitulate human cancer have proven indispensable for studying aspects of metastasis that have remained unclear until now.
For example, metastasis was originally believed to be a late step in tumorigenesis.
However, against all expectations, studies in BALB-NeuT and MMTV-PyMT mouse mammary tumor models revealed that transformed cells in early lesions are already capable of disseminating to bone marrow and lungs to form micro-metastasis 106.
Figure 1. The utility of mouse models in metastasis research
This overview summarizes the utility of different preclinical mouse models of experimental and spontaneous metastasis to study the different steps of the metastatic cascade. Conventional GEMMs represent oncomice and mice carrying germline mutations in TSGs. Next-generation GEMMs represent mouse models that are genetically engineered to accurately mimic sporadic human cancer. For some models, the utility for studying specific steps in the metastatic cascade has yet to be determined, as indicated by a question mark. Moreover, several studies have shown that components of the adaptive immune system contribute to the various steps of the metastatic cascade. These aspects cannot be studied in models based on xenografting of human cancer cells or tumor fragments in immunodeficient hosts (indicated by an asterisk). To circumvent this, humanized mice can be used as hosts.
Similarly, epithelial-to-mesenchymal transition (EMT) – a process in which cells lose their polarity and cell-cell adhesion, and gain migratory properties – is thought to play a key role in tumor cell dissemination and metastasis. Using a spontaneous squamous cell carcinoma mouse model it was found that reversible EMT, regulated by spatiotemporal expression of Twist1, is essential for metastasis formation 107. However, recent studies in GEMMs of pancreatic and breast cancer show that cancer cells retain their epithelial characteristics whilst colonizing metastatic sites, suggesting that EMT is not essential for metastasis formation in these models 108,109. Together these studies emphasize the complexity of spontaneous metastasis.
GEMMs have also revolutionized the metastasis field by revealing complex crosstalk between cancer cells and the immune system in metastasis formation. Several labs have shown that myeloid immune cells, such as macrophages and neutrophils, play key roles in promoting metastasis formation in different types of cancer 100,110–113. Recently, we reported a mammary tumor-induced systemic inflammatory state characterized by IL17- producing γδ T cells and the subsequent expansion of immunosuppressive neutrophils that drives spontaneous metastasis formation in a GEMM of lobular breast cancer and a GEMM-based transplantation model for spontaneous metastatic disease 112. Collectively, GEMMs have proven indispensable for understanding the complexity of metastasis and have challenged the current dogma that metastasis is a late-stage cancer cell-intrinsic process involving EMT. These findings may have important implications for treatment of metastatic cancer patients.
Applications of GEMMs in translational oncology
Besides providing essential insights into basic cancer research, GEMMs that harbor patient- relevant allelic variants of known cancer genes have proven detrimental for translation oncology. Close alignment of mouse and human studies can provide a platform that can aid in the development of novel treatment strategies for cancer patients (Figure 2).
Below we discuss how GEMMs can provide clinically relevant information on the design of anti-cancer therapy.
Validation of novel drug targets
Considering that not all cancer genes are essential for maintenance of established tumors, it is important to test whether reactivation of a TSG or down-regulation of an oncogene results in durable regression of established tumors in a realistic preclinical setting, before drugs against these targets are developed. The relevance of oncogenes for tumor maintenance can be assessed in inducible mouse models in which oncogene expression can be de-induced once tumors have developed. For example, de-induction of oncogenic Pik3caH1047R expression in a mouse model of breast cancer caused (partial) tumor regression demonstrating that these tumors are ‘addicted’ to activated PI3K signaling. However, most tumors eventually recurred due to Met or Myc amplifications, indicating that these genetic lesions may induce resistance to PI3K inhibitors 114. This
2
example illustrates that preclinical studies in inducible GEMMs are not only useful for validating drug targets but also for identifying mechanisms underlying acquired drug resistance.
Also TSGs may in certain cases constitute valid drug targets. For example, p53 loss- of-function in cancer can result from dominant-negative or inactivating mutations in the Trp53 gene or from amplification/overexpression of its specific inhibitors MDM2 and MDM4. Genetic studies in GEMMs with reversible inactivation of p53 have shown that restoration of p53 leads to rapid regression of established tumors 44,115,116, providing strong rationale for designing anticancer drugs that restore p53 function by inhibiting MDM2 117 or by restoring wild-type function to mutant p53 118. Similarly, GEMMs of colorectal cancer with inducible knockdown of APC showed that APC restoration initiates rapid and extensive tumor cell differentiation and sustained regression without relapse, providing in vivo validation of the WNT pathway as a therapeutic target for treatment of APC-mutant colorectal cancers 59.
Kerstenet al., Figure 2
Figure 2. The utility of mouse models in cancer drug development
Development of novel treatment strategies in oncology requires preclinical studies in mouse cancer models to identify and validate novel cancer drivers and therapeutic targets, to determine in vivo drug pharmacokinetics and pharmacodynamics (PK/PD), and to evaluate in vivo anti-cancer efficacy of novel therapeutics. When promising preclinical results are obtained, the tolerability and anti-cancer efficacy of these drugs are evaluated in human patients in phase I-III clinical trials. A fraction of patients will show poor response due to intrinsic or acquired resistance, which may be studied mechanistically in preclinical mouse models to identify response biomarkers and combination therapies to prevent or overcome resistance. The close alignment of mouse studies and human clinical trials will lead to better patient stratification, identification of novel biomarkers and development of optimal combination therapies, culminating in improved cancer patient care.
Unraveling therapy response and resistance
To minimize the risk of failure of novel anti-cancer therapeutics in clinical trials, preclinical evaluation of response and resistance in robust and predictive in vivo models is essential. Therapeutic responses of GEMMs to targeted therapy and conventional chemotherapy are very similar to those of human patients, as was assessed in GEMMs of Kras-mutant lung cancer and pancreatic cancer 119. Hence, preclinical drug efficacy studies in GEMMs may advance the development of optimal (combinations of) anti- cancer drugs to target specific tumors, and the identification of determinants of therapy response that may be used as predictive biomarkers for patient stratification.
In addition, GEMMs may be used to identify mechanisms by which therapy-sensitive tumors acquire drug resistance.
A clear example of a preclinical GEMM that has provided mechanistic insight into therapy response and resistance of BRCA1-mutated breast cancer is the K14cre;Brca1F/F; Trp53F/F (KB1P) mouse model. KB1P mice develop mammary tumors that mimic the histopathological features of human BRCA1-mutated breast cancers as well as their hypersensitivity to platinum drugs and PARP inhibitors 120,121. Clinical trials evaluated the PARP inhibitor olaparib for the treatment of ovarian, breast and colorectal cancer 122. While olaparib did not seem promising in this diverse group of cancer patients, it did show significant responses in BRCA1-mutation carriers, due to the synthetic lethal combination of PARP inhibition and BRCA1-deficiency 123,124. BRCA1-mutant cells are more vulnerable to PARP inhibition because the single-strand DNA breaks induced by PARP inhibition, lead to double-strand breaks during replication, which cannot be repaired by BRCA1-deficient cells due to lack of homologous recombination. Based on promising results obtained in clinical trials 123,124, olaparib (trade name LynParza) was FDA approved in December 2014 for the treatment of patients with advanced BRCA1/2- mutated ovarian cancer. Despite the good response of BRCA1/2-mutated cancers to olaparib, acquired resistance is observed both in patients and GEMMs. Preclinical studies in KB1P mice revealed several mechanisms of resistance, such as elevated levels of drug efflux transporters and restoration of homologous recombination 121,125–127. These studies could aid in understanding clinical resistance and in designing improved treatment strategies for olaparib-resistant patients in the clinic.
It is becoming clear that therapy response and resistance is not only influenced by tumor cell-intrinsic factors but also by stromal factors such as fibroblasts and immune cells 128–132. The impact of these tumor cell-extrinsic factors can be more effectively studied in GEMMs than in xenograft models, as GEMMs closely recapitulate the constant crosstalk between cancer cells and their natural microenvironment. This is illustrated by tumor intervention studies in a GEMM of PDAC, which showed that therapeutic inhibition of paracrine Sonic Hedgehog (SHH) signaling reduced desmoplastic tumor stroma and increased tumor vasculature, resulting in enhanced delivery of gemcitabine to tumors 133. However, the concept of targeting tumor stroma in PDAC has recently been challenged by two studies showing that stromal factors may suppress rather than promote PDAC growth, possibly by restraining tumor angiogenesis 96,134. Together, these
2
studies demonstrate that the contribution of the tumor microenvironment to therapy resistance may be more profound but also more complex than previously anticipated.
Cancer immunotherapy
Over the past decades, increasing mechanistic insights into the principles of immune responses have culminated in therapeutic strategies that harness the patient’s immune system to attack cancer. Recent clinical trials in patients with advanced melanoma and lung cancer confirm the remarkable potential of immune checkpoint blockade, including anti-CTLA-4 and anti-PD-1, to enhance effective anti-tumor immunity and to improve survival in a proportion of the patients 135,136. The basis of these clinical trials comes from several decades of fundamental research in experimental mouse models that have revealed the importance of CTLA-4 and PD-1 in restraining immune responses, as most clearly illustrated by the severe spontaneous autoimmunity phenotype in CTLA-4-deficient 137 and to a milder extent in PD-1-deficient mice 138,139. A seminal study from Allison and colleagues showed that CTLA-4 blockade in mice bearing inoculated tumors enhances anti-tumor T cell responses resulting in tumor rejection 140, illustrating that releasing the brake on T cells might be an interesting strategy to combat cancer.
Nevertheless, a substantial proportion of patients do not respond to immunotherapy, and the current challenge is to understand why.
Although the majority of immunological studies are performed in transplantation models, we foresee a growing role for GEMMs that closely mimic human cancer patients in terms of genetic drivers, tumor histopathology and the crosstalk between cancer and immune cells that co-evolve with the developing cancer. Several studies in GEMMs show that during de novo carcinogenesis, tumor-specific T cell responses are dysfunctional due to tumor-induced tolerance mechanisms 141–143. However, transplantation of GEMM-derived tumor cells in immunodeficient mice resulted in rapid tumor growth, while wild-type mice rejected these tumors 141–143, demonstrating that these tumor cells did not lose their immunogenicity and T cells are still able to recognize and attack them.
Why anti-tumor T cell responses fail to control de novo tumors remains largely unclear.
Many tumors are characterized by chronic inflammation, which induces local and systemic immunosuppression that is unfavorable for T cells to perform their effector function 112,144,145. Moreover, tumors often show dysfunctional dendritic cells, which results in impaired T cell priming. In the MMTV-PyMT mammary tumor model a rare population of dendritic cells can be found that are very potent activators of anti-tumor T cells 146. However, these cells are outcompeted by the overabundant presence of macrophages preventing proper T cell activation 146,147. Recent studies have demonstrated that boosting dendritic cell function 144,146,148,149 or blocking myeloid cell-induced immunosuppression 150,151 improves the anti-tumor efficacy of immune checkpoint blockade. Thus, patients that show acquired resistance to T cell boosting immunotherapy might show improved clinical benefit when treatment is combined with compounds that either target immunosuppression or enhance T cell priming.
